I think we need Coulomb to comment on this.
Most if it has been said pretty well.
I've softened my stance a little on charging LFP to 100%. Degradation is all about cell voltage and temperature; both encourage reactions, good and bad. So hot battery cells perform better, but degrade faster. But the damage depends on the time at the high temperature and voltage. That's why Teslas will preheat a battery to get the highest fast charge rate; because it's short duration, it won't harm the battery significantly.
Temperature is usually hard to avoid, though if you can park your car in the shade in summer, it's worth doing that.
LFP cell voltage is already about 15% lower than that of NMC, going by either nominal voltage (3.2 versus 3.75 V, or fully charged voltage (3.6 versus 4.2 V). That's why overall, LFP lasts longer. It lasts more than 15% longer because it's a sort of exponential effect.
LFP has one more trick: once you fully charge the cell to 3.6 V (sometimes 3.65 V), the voltage relaxes to about 3.5 V over an hour or so, so that's another 3% advantage in resting voltage. This relaxation of terminal voltage doesn't seem to cause significant reduction in stored energy. If you use the car right away after charging, the extra voltage disappears very soon anyway, so charging to the highest voltage doesn't store much extra energy. So if you're really paranoid about battery life, you could try to charge to say 98% max. Sadly, MG doesn't help you with this, as they disable the charge SoC limit on the SR models. I don't know if charging to 98% is enough to reset the BMS SoC guesstimate; it should be, because by then the cell should be on one of the steep parts of the voltage versus SoC curve.